{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/source-primed-multi-turn-conversation-helps","title":"Source-primed Multi-turn Conversation Helps Large Language Models Translate Documents","arxiv_id":"2503.10494","date":"2025-03-13","proceeding":null,"authors":["Hanxu Hu","Jannis Vamvas","Rico Sennrich"],"abstract":"LLMs have paved the way for truly simple document-level machine translation, but challenges such as omission errors remain. In this paper, we study a simple method for handling document-level machine translation, by leveraging previous contexts in a multi-turn conversational manner. Specifically, by decomposing documents into segments and iteratively translating them while maintaining previous turns, this method ensures coherent translations without additional training, and can fully re-use the KV cache of previous turns thus minimizing computational overhead. We further propose a `source-primed' method that first provides the whole source document before multi-turn translation. We empirically show this multi-turn method outperforms both translating entire documents in a single turn and translating each segment independently according to multiple automatic metrics in representative LLMs, establishing a strong baseline for document-level translation using LLMs.","url_abs":"https://arxiv.org/abs/2503.10494v1","url_pdf":"https://arxiv.org/pdf/2503.10494v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"source-primed-multi-turn-conversation-helps","repo_url":"https://github.com/ZurichNLP/multiturn-llm-docmt","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"document-level-machine-translation","task_name":"Document Level Machine Translation"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2503.10494","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}